Photoelectron spectra of the rhenium pentacarbonyl halides

crease was the result of increasing ionic strength due to the. SCN' was ruled out on the basis of the following experiment. The kinetic data for [N02'...
0 downloads 0 Views 433KB Size
Inorganic Chemistry, Vol. 13, No. 12, I974 3035

Notes In other words the rate constant should decrease with increasing nucleophile concentration. In addition kobsd should increase with increasing nitrite concentration. This behavior was observed. For [NO2-] fixed at 0.025 M , kobsd decreases from 6.3 (at [SCN-] = 0.02 M ) to 4.9 (at [SCN-] =0.1 M ) . For [NO,-] = 0.1 M , the k,,bsd values are higher but still decrease with increasing [SCN-], from 10.1 ([SCN-] = 0.02 M ) to 5.9 ([SCN-] = 0.1 M ) . The possibility that this rate decrease was the result of increasing ionic strength due to the SCN- was ruled out on the basis of the following experiment. The kinetic data for [NO2-] = 0.1 M were collected at two ionic strengths, 0.25 M (data above) and 1 .OM (KNO,). The higher ionic strength caused not a decrease but an increase in kobsd of about 40%. Thus the decrease in rate with increasing [SCN-] cannot be an ionic strength effect. We feel that our results are in good agreement with that predicted from the data of Blanck, et al.,' especially when small differences in ionic strength and temperature are considered. The conclusion is that the rates of ligand substitution on ferric myoglobin can be quantitatively predicted from a knowledge of aquation and anation rates. A species such as Y-Mb-L never becomes important during the reaction. For anation and aquation one expects the usual dissociative interchange mechanism to hold and the results for ligand substitution presented herein clearly support this. Acknowledgment. Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for the support of this research. Registry No. Horse heart myoglobin, 12585-26-1; Fe3+,2007452-6; SCN-, 302-04-5; N 3 - , 14343.69-2; OCN-, 661-20-1; NO,-, 14797-65-0; CN-, 57-12-5.

Contribution from the Departments of Chemistry, Texas A&M University, College Station, Texas 77843, and the University of Rochester, Rochester, New York 14627

Photoelectron Spectra of the Rhenium Pentacarbonyl Halides Gerald P. Ceasar,*la Paul Milazzo,la John L. Cihonski,'b>c and Robert A. Levenson*'b

Received May 28, 19 74

AIC403399

He1 ultraviolet photoelectron spectroscopy (UPS) of the manganese pentacarbonyl halides Mn(CO),X (X = C1, Br, I) has recently received considerable attention?-5 From experimentalZasand theoretical33 4 considerations it has been shown that, in contrast to the situation normally observed in transition metal complexes, the highest occupied molecular orbital is primarily of ligand character. The most recent interpretation of U P S is that of Lichtenberger, Sarapu, and Fenske, who proposed, partly on the basis of theoretical calculations, that the energy ordering of the four highest filled (1) (a) University of Rochester. (b) Texas A&M University. (c) This research constitutes partial fulfillment of the Ph.D. degree requirements of J. L. C. (2) D. L. Lichtenberger, A. C. Sarapu, and R. F. Fenske, Znovg.

Chem.. 12 702 (1973). (3) R. F. Fenske and R. L. DeKock,Znorg. Chem., 9, 1053 (1970). (4) D. A. Brown and W. J. Chambers,J. Chem. Soc. A , 2083 (1971). (5) S. Evans, J. C. Green, M. L. H. Green, A. F. Orchard, and D. W. Turner, Discuss. Faraday Soc., 47, 112 (1969).

molecular orbitals is X, > Xu > e(d,,, dy,) > b,(d,,).* The X, and Xu levels are of symmetry e and a l , respectively, but the X, and Xu notation will be used exclusively in this paper. The present paper describes our investigations of the lowenergy photoelectron spectra of the Re(CO)SX series. This study was of interest since UPS has been used in a relatively limited number of transition metal complexes with noncubic symmetry and since an interpretation of the electronic spectra of the Mn(CO),X species has been based on the orbital ordering obtained from U P S results.6 By contrasting the ionization potentials obtained for the Re series with those of the Mn complexes, we wished to determine the changes which take place in orbital structure with variation of the central metal atom and to see whether these data are consistent with previous interpretations. Our results now lead us to propose a new energy level ordering which we believe is appropriate for both the Mn and Re complexes. Experimental Section The rhenium pentacarbonyl halides were synthesized by methods analogous t o those used t o prepare the corresponding Mn comp l e x e ~ . ' ~He1 ~ spectra were recorded on a Vacuum Generators ESCA2 photoelectron spectrometer operating a t 35-MeV resolution. The iodide complex was sufficiently volatile to allow gas-phase spectra t o be obtained at room temperature. For Re(CO),Cl and Re(CO),Br it was necessary t o heat the ionization chamber t o 70" before spectra could be obtained,

Results The ultraviolet photoelectron spectra of the Re(CO),X series are shown in Figure 1 for the critically important 8-12eV region and compared with the results on the Mn analogs2 in Table I. (In constructing Table I, the position of bands split by spin-orbit coupling is taken as the average of the two components.) For the purpose of the discussion which follows ionization potentials are labeled I-IV with the lowest ionization being I. The Re(CO),X spectra are characterized by several features which distinguish them from those of the Mn complexes. First, band I1 is seen for all X (Cl, Br, I) whereas it is not directly observed in Mn(CO)51. Second, band I, assigned as an ionization from a degenerate level on intensity grounds: exhibits more pronounced spin-orbit splitting in the case of Re(CO),I than in Mn(CO),I and even shows splitting for Re(CO)SBr whereas none is seen in Mn(CO)5Br. These results suggest that the highest occupied molecular orbital, which we assign as being primarily of halogen parentage (vide infra), has a not insignificant amount of metal character. The presence of metal character in this level would then increase the observed spin-orbit splitting since the spin-orbit coupling constant for Re is much larger than that of Mn (0.3 vs. 0.03 eV). An alternative explanation for the larger splitting in the Re complexes is that the X, level is more nearly pure halogen in character than in the Mn species. Third, band I11 for Re(CO),Br is split into two peaks separated by 0.12 eV while the chloride and iodide show no apparent splitting for this level. According to the arguments made above, band 111, which is primarily of metal character (vide infra), should show some doublet splitting. Its absence in the chloride and iodide derivatives may result from changes in the amount of metal character due to halogen or carbonyl 71 mixing' or possibly from inductive changes in the effective charge on the metal, since the spin-orbit coupling (6) G. B. Blakney and W. F. Allen, Znorg. Chem., 10,2763 (1971). (7) E. W. Abel and G. Wilkinson, J. Chem. Soc., 1501 (1959). (8) E. 0.Brimm, M. A. Lynch, Jr., and W. J . Sesny, J. Amer. Chem. Soc., 76, 3832 (1954).

Notes

3036 Inorganic Chemistiy, Vol, 13, No. 12, 1974 10.82

I

8.0

9.0

10.0

11.0

12.0

ev.

Figure 1. Ultraviolet photoelectron spectra of the metal d and halogen p regions for Re(CO),X. Table I. Ionization Potentials of the Pentacarbonyl Halides of Mn and Re (eV) Band Mn(GO),Xa Re(CO),Xb DiffC I I1

x = c1

IV

8.94 9.56 10.56 11.18

I I1 III IV

8.86 9.56 10.11 10.81

111

9.02 9.95 10.82 11.22

0.08 0.39 0.26 0.04

8.86 9.94 10.46 10.85

0.0 0.38 0.35 0.04

X = Br

X=I I I1 111 1v a

8.59

8.50 9.77 10.06 10.47

9.69 10.44

Reference 2. b This work.

C

-0.09 0.34 0.03

IP(Re) - IP(Mn).

constant is dependent on this parameter. In any case, the question of the origin of the splitting of band I11 will not materially affect the remaining discussion. Discussion We propose that the photoelectron spectra are best interpreted assuming the level ordering X, > b2 > e > X,. This corresponds to an inversion of the positions of the nondegenerate b2 and X, levels in the scheme suggested by Lichtenberger, Sarapu, and Fenske.2 The basis for this conclusion comes from a comparison of the photoelectron data

for t.he Mn and Re complexes with each other and with other halogen-containing compounds. First, from Table I it can be seen that bands I and XV are functions of X only and are virtually independent of M while bands I1 and III shift to higher IP by ca. 0.4 eV as M changes from Mn to Re. A change of this magnitude is consistent with the first IP for We being 0.4 eV greater than that of Mn and with the observation that %%emetal IP's are shifted ca. +0.4 eV in compounds like compared to their 1Mn analogs. These correlations clearly pinpoint IK and 11%as b2 and e metal type levels and I and XV as ionizations from the X, and X, orbitals. Support for this assignment comes from the fact that band I1 varies least with X for both the Mn and Re series. This is precisely what is to be expected for the b2 metal carbonyl level. This orbital is nonbonding with respect to the halogerr and should change only slightly due to inductive effects as the halogen is changed. It should, however, be sensitive to the central metal atom as is shown by the IP data. Halogen levels, on the other hand, would be expected to vary as X is changed, as would the metal e orbital to some extent since it has the proper symmetry t o mix with the X, orbital. Band IV, which does vary with X,would thus seem to be a poor choice for the b2 level on the basis of the data in Table 1. It should also be pointed out that the IP o f IH is in good agreement with energy values we estimate for the bZ level. This was done by using electronic spectral work of Beach and Gray'' which established that the 2tzu n* CO level in hexacarbonyl complexes is essentially invariant to the central metal. Their analysis placed the TC*level at a constarit 23,400 i 600 cm-' for Cr(CO)6, Mo(CO),, and W(GQ16. Using this value and the energy of the 2t2g-3 2tzu transition for Mn(GO),+ and Re(C0)6' we predict the b 2 level to be at ca. 9.1 and 9.3 eV for the present Mn and Re complexes. These numbers are in reasonably good agreement with the observed trend in experimental IP's for band 11. Simple molecular orbital theory also helps in ordering the metal levels once it has been established that the highest filled level corresponds to the X, level sf e symmetry. To first order, on descending jn symmetry from 0,to C4"$the eg orbitals are split by u interactions while the t2glevel is split by n interactions, If we consider a M(C ),X moleciale with the X, level above the initially unsplit b2 and e metal levels, simple molecular orbital theory predicts that the e level should drop below the noninteracting b2 level and not be above it; this is not an unreasonable explanation since the b2 level should be unaffected b y the X group. We note that the assignments of Eichtenberger, et al., re quire that bands I11 and IV correspond to meta.1 levels. Accepting this assignment thus means eit bz-e ligand field separation exists for the this separation decreases in the Mn complex in tlie ortler I > Br > Cl. Our assignment indicates that the b 2 -e separation decreases in the order C l > Br > I exactly as would be suggested by the spectrochemical series. Further evidence for our ordering i s gained by observing that the energy difference between bands I and 11 for the Re series is 0.90, I .lo, and 1.20 eV for X = C1, respectively. This increasing energy difference is opposite to the decreasing X,-X, separation which is generally seen in (9) 6.P. Ceasar, P. Milazzo, J . L. Cihonski, and R. A. Levenson, unpublished results. (10) S. Cradock, E. A. V. Ebsworth, and A. Robertson, J~Chenz. SOC.,Dalton Trans., 22 (1973). (1 1) N. A. Beach and H.B. Gray, J . Amer. C h ~ mS. O C . ,90,5 7 1 3 ( 1 968).

Inorganic Chemistry, Vol. 13, No. 12, 1974 3037

Notes halogen compounds on going from C1 to Br t o I (e.g., CH&,12 HX,12 X2”). Thus, our assignment of band IV as an ionization from the X level leads to a more acceptable X,-Xu separation with respect to the trend on variation of halogen. We believe that our results and line of reasoning (applicable to both the Mn and Re complexes) clearly establish the ordering X, > b2 > e > Xu for the M(CO)5X molecules under discussion. The advantages of varying the metal as well as the ligand as an aid to interpreting photoelectron spectra of transition metal complexes have been clearly demonstrated. Acknowledgments. We thank the Robert A . Welch Foundation (Grant A-509), the Office of Naval Research, and the donors of the Petroleum Research Fund, administered by the American Chemical Society, for support of this research. Registry No. Re(CO),Cl, 14099-01-5; Re(CO),Br, 14220-21-4; Re(C0) ,I, 13821-00-6. (32) D. W. Turner, C. Baker, A . D. Baker, and C. R. Brundle,

“Molecular Photoelectron Spectroscopy,” Wiley-Interscience, London, 1970, Chapter 8. (13) D. C. Frost, C. A. McDowell, and D. A. Vroom, J. Chem. Phys., 46, 4 2 5 5 (1967).

NU 2 0 1 4 SN TPP CL2

$ e o

* P-

0

W



5% t--

z 74

rL1

>

p -

0

d

a

0

0

CHFINNELS Figure 1. ’lgmSn Mossbauer spectrum of dichlorotetraphenylporphinatotin(IV), Cl,SnTPP, a t 80 K.

Table I. ‘lgmSn Mossbauer Spectral Parameters of Para-Substituted C1,SnTPP -

Contribution from the Department of Chemistry, Colorado Women’s College, Denver, Colorado 80220 and New England Institute, Ridgefield, Connecticut 06877

Tin-1 19m Mossbauer Spectra of ParaSubstituted Dichlorotetraphenylporphinatotin(1V) Complexes Nordulf W. G. Debye* and Alan D. Adler

Received June 4, 1974

AIC40 359 A

In a previous report of the Mossbauer spectra of various tin complexes of phthalocyanine and tetraarylporphins,’ Curran and O’Rourke indicated their inability to resolve a possible quadrupolar splitting in the spectra of four dichlorotetraarylporphinatotin(1V) compounds. We now wish to report the successful resolution of the quadrupolar splittings in seven such compounds including those studied by Curran and O’Rourke. The compounds are all derivatives of dichlorotetraphenylporphinatotin(IV), hereafter identified as C12SnTPP, differing by various substitutions at the para positions of the four phenyl groups. The seven substituents studied were F , C1, Br, CH3, OCH3, and iC3H7,in addition t o the H of the parent compound, and were prepared and characterized by previously reported methods? A representative spectrum clearly showing such a resolved splitting is shown in Figure 1 and a summary of all the llhllSn Mossbauer spectral parameters obtained at 80 K is given in Table I. All spectra were obtained with a Ba’lgmSn03 source held at room temperature and the isomer shift (IS) values are reported with respect to a BaSnO, absorber at room temperature. The uncertainties indicated in parentheses for both the IS and the quadrupole splitting (QS) represent the lo: con-

Compd Cl,SnT(p-CC,H,)PP Cl,SnT(pCH,)PP C1, SnTPP Cl,SnT(p-OCH,)PP Cl, SnT( p-Br)PP Cl, SnT(pC1)PP C1, SnT( p-F)PP

1sa-c

0.160 (0.010) 0.153 (0.008) 0.146 (0.007) 0.147 (0.008) 0.147 (0.005) 0.154 (0.006) 0.165 (0.013)

Area O S ~ J ’ ? ~ ratioe

0.512 (0.012) 0.592 (0.010) 0.588 (0.008) 0.555 (0.010) 0.659 (0.007) 0.642 (0.007) 0.463 (0.018)

0.65 0.77 0.73 0.75 0.80 0.87 0.60

a Isomer shift (with respect to BaSnO,) and quadrupole splitting in mmlsec, defining the room-temperature parameters of our (CH,),SnF, calibration sample to be IS = 1.2455 mmlsec and QS = 4.4945 mmlsec. All spectra were recorded at 80 K. b Numbers in parentheses represent the ca. 68% confidence level in the precision. C Correlation index for the two wing positions taken t o be +1.0 during calculation of u. d Correlation index for the two wing positions taken t o be 0.0 during calculation of u. e Area of high-energy wing divided by area of low-energy wing.

* To whom correspondence should be addressed at Colorado Women’s College. (1) M. O‘Rourke and C. Curran, J. Amer. Chem. Soc., 9 2 , 1501

fidence range for the precision of the evaluated parameters; we estimate the overall la confidence level, including both uncertainties in the precision and the accuracy, t o require the addition of 0.01 mmlsec to the stated numbers. The total count accumulation per data point for these spectra range from one to ten million and all compounds demonstrated a large resonance effect. Spectrometer calibration spectra were “time-sandwiched’’ about each sample spectrum and the expressed uncertainties are the pooled uncertainties of each sample spectrum and two calibration spectra. The spectral parameters were obtained with the aid of the PARLOR computer program3 which assumes that the spectra represent independent lorentzian absorption curves on a parabolic base line. The seven compounds clearly exhibit variations in the quadrupole splitting outside of the stated uncertainties in addition to more poorly defined variations in their isomer shifts. The presence of these resolvable splittings is somewhat surprising as the tin atom is bonded only to other atoms which all possess lone-pair electrons, a bonding situation which usually results in a singlet spectrum. Correlations of both spectral parameters (IS, QS) with the following empirical

(1970). (2) A. D. Adler, F. R. Longo, J . D. Finarelli, J. Goldmacher, J . Assour, and L. Korsakoff, J. Org. Chem., 3 2 , 476 (1967); A. D. Adler, F. R. Longo, F. Kampas, and J . Kim, J. Inorg. Nucl. Chem., 3 2 , 2443 (1970); A. D. Adler and J . L. Harris, Anal. Biochem., 14, 472 (1966).

( 3 ) E. Rhodes, W. O’Neal, and J. J . Spijkerman, Nut. Bur. Stand. (U. S.), Tech. Note, No. 104, 108 (1966); further information concerning this program may be obtained from Dr. John C. Travis, National Bureau of Standards, Washington, D. C. 20234.